Engineering Structures 33 (2011) 621634

Contents lists available at ScienceDirect

Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct

Evaluation of drift demands in existing steel frames under as-recorded far-field

and near-fault mainshockaftershock seismic sequences
Jorge Ruiz-Garca , Juan C. Negrete-Manriquez
Facultad de Ingeniera Civil, Universidad Michoacana de San Nicols de Hidalgo, Edificio C, Planta Baja, Cd. Universitaria, 58040 Morelia, Mexico

article info abstract

Article history: This paper presents results of a study aimed at evaluating the effect of aftershocks in steel framed
Received 20 July 2010 buildings. For that purpose, three frame models representing existing steel moment-resisting frames
Received in revised form were subjected to a set of as-recorded mainshockaftershock seismic sequences. For this purpose, 64
13 October 2010
as-recorded seismic sequences registered as a consequence of the 1994 Northridge and 1980 Mammoth
Accepted 10 November 2010
Available online 13 December 2010
Lakes earthquakes were considered in this study. In particular, this investigation employed 14 seismic
sequences recorded in 7 accelerographic stations in the near-fault region. An examination of the as-
Keywords:
recorded seismic sequences shows that the frequency content of the mainshock and the main aftershock
Aftershocks is weakly correlated. The response of the frame models was measured in terms of the peak and residual
Seismic sequences (permanent) drift demands at the end of the earthquakes excitation. From the results of this investigation,
Artificial seismic sequences unlike previous results based on artificial seismic sequences, it was found that as-recorded aftershocks do
Peak drift demands not significantly increase peak and residual drift demands since the predominant period of the aftershocks
Residual drift demands (i.e. frequency content) is very different from the period of vibration of the frame models. In addition, it
was shown that artificial seismic sequences could significantly overestimate median peak and residual
drift demands as well as the record-to-record variability.
2010 Elsevier Ltd. All rights reserved.

1. Introduction because of the technical difficulties in straightening and repairing

There is a consensus among the earthquake engineering
community that damage in structural elements, and some drift-
sensitive nonstructural components, is primarily the result of lat-
eral deformation demands induced by earthquake ground shaking
in the structure. As a consequence, modern performance-based
assessment methodologies for evaluation of existing structures
(e.g. FEMA 356 [1]) are based on the estimation of peak lateral dis-
placement demands that man-made structures could suffer under
seismic excitation. However, man-made structures located in seis-
mic regions are not exposed to a single seismic event (i.e. main-
shock), but also to a seismic sequence consisting of foreshocks,
mainshock and aftershocks. For example, after the mainshock on
February 27, 2010 (Mw = 8.8) that struck the central-southern re-
gion of Chile, 306 aftershocks having magnitudes greater than 5.0
were recorded between February 27 and April 26. Among them,
21 aftershocks had magnitudes greater than 6.0. In particular, his-
torical earthquakes have shown that the aftershocks may increase
the damage state at the end of the mainshock. For example, sev-
eral dozen damaged reinforced concrete (RC) buildings in Mexico
City had to be demolished after the September 1985 earthquakes
Corresponding author. Tel.: +52 443 2278545; fax: +52 443 3221002.
E-mail address: jruizgar@stanfordalumni.org (J. Ruiz-Garca).
0141-0296/$ see front matter 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.engstruct.2010.11.021
buildings with large permanent drifts and the future threat due
to aftershocks [2]. Another example of the consequence of signif-
icant permanent lateral displacements is a two-story steel office
building that was severely damaged during the 1994 Northridge
earthquake (Mw = 6.7) [3]. The excessive permanent displace-
ments were a consequence of concentrated structural damage in
the first story, with yielding noted at the base plate connections,
and a number of column fractures at the second floor moment
connections. The extent of this damage was such that the build-
ings owner decided to demolish the structure above the founda-
tion level. However, earthquake field reconnaissance reports af-
ter the 1994 Northridge earthquake mentioned that other existing
steel buildings could have experienced permanent lateral defor-
mations as a consequence of the strong mainshock that affected
the Los Angeles area [4], but they did not report that the follow-
ing aftershocks could have increased the level of damage, or even
to drive damaged buildings to the risk of collapse. Therefore, there
is a need to further understand the effects of as-recorded main-
shockaftershock seismic sequences in the seismic response of ex-
isting structures.
There have been several investigations aimed at studying the ef-
fect of seismic sequences on the response of civil engineering struc-
tures [515]; some of them have been focused on the nonlinear re-
sponse of single-degree-of-freedom (SDOF) systems [510] while
622 J. Ruiz-Garca, J.C. Negrete-Manriquez / Engineering Structures 33 (2011) 621634
others in the response of multiple-degree-of-freedom (MDOF) sys-
tems [1115]. A pioneering analytical study of nonlinear SDOF sys-
tems subjected to mainshockaftershock acceleration time histo-
ries recorded during the 1972 Managua earthquake was carried
out by Mahin [5]. He observed that the displacement ductility de-
mand, (i.e. peak inelastic displacement normalized with respect
to the systems yield displacement) of elastoplastic SDOF systems
slightly increased at the end of the main aftershock with respect to
the mainshock.
It should be noted that few studies have focused their attention
on the nonlinear response of MDOF systems subjected to seismic
sequences [1115]. For example, Lee and Foutch [12] as well as Li
and Ellingwood [13] studied the response of two steel moment-
resisting frame models having 9- and 20-story heights designed
to be representative of typical office building construction in
the Los Angeles area prior to the 1994 Northridge earthquake
(Mw = 6.7) when subjected to artificial seismic sequences. Both
studies employed earthquake ground motions developed during
the SAC project [16] that represent specific hazard levels (e.g. 2%
and 10% probability of exceedance in 50 years), as mainshocks.
For instance, Li and Ellingwood [13] employed a methodology
proposed by Sunasaka and Kiremidjian [6] to generate artificial
aftershocks by scaling the mainshock by a factor derived from
the aftershock hazard near Eureka, California. For example, the
scale factor is 0.68 when moment magnitudes of 6.8 and 6.2
are assumed for the mainshock and aftershock, respectively,
while a scale factor of 1.0 is adopted when equal mainshock
and aftershock magnitudes are assumed. In addition, they also
derived a set of sequences by randomizing the ensemble of
mainshocks and generated aftershocks in order to simulate
the stochastic dependence between them. From this study,
they found that the approach of repeating the mainshock as
an aftershock tends to overestimate the peak inter-story drift
demands in the analyzed frames as compared to the randomized
approach. Later, Ruiz-Garca et al. [14] examined the response of
9 typical low-height reinforced concrete highway bridges under
26 as-recorded mainshockaftershocks sequences gathered in the
subduction zone of the Mexican Pacific Coast. They found that
aftershocks do not significantly increase drift demands due to
the inherent overstrength in the study-case low-height highway
bridges. Using incremental dynamic analysis (i.e. scaling each
mainshockaftershock sequence to reach different levels of ground
motion intensity), they found that the effect of aftershocks tends
to increase both peak and residual (permanent) drift demands
when the bridge models behave nonlinearly during the mainshock.
The increment in drift demands depends on the level of the
ground motions intensity and the type of hysteretic behavior
considered in the bridge columns. More recently, Hatzigeorgiou
and Liolios [15] studied the response of 4 regular and 4 irregular
reinforced concrete frames under five as-recorded and 40 artificial
seismic sequences. They concluded that the frames consistently
increased their displacement ductility demands under the five real
seismic sequences.
It should be noted that although previous studies provided in-
formation on the effect of seismic sequences on the response of
structures, most of the previous studies employed artificial seis-
mic sequences instead of as-recorded mainshockaftershocks se-
quences to evaluate the seismic response [e.g. [613,15]]. They
employed artificial sequences using the mainshock acceleration
time-history as a seed for simulating the following aftershocks, ei-
ther repeating (i.e. back-to back approach, [7,9,1113,15]) or ran-
domizing the set of mainshocks [e.g. [8,10,13]]. The first approach
assumes that the ground motions features such as amplitude, fre-
quency content, and strong motion duration of the mainshock and
aftershock(s) are the same. However, this simulation procedure
is the worst seismic scenario which might be unlikely to occur
in earthquake-prone regions [13]. In the absence of as-recorded
seismic sequences, the second approach could be more appropri-
ate for reproducing a mainshockaftershock seismic environment,
but it has not been evaluated against results from as-recorded se-
quences. Thus, there is still a need of investigating the response
of structures under real (i.e. as-recorded) mainshockaftershock
sequences. Available strong motion databases, such as the Pacific
Earthquake Engineering NGA Database (PEER NGA) [17], provide
a good opportunity for employing acceleration time-histories of
recorded mainshocks and their corresponding aftershocks.
The main objective of this paper is to gain further understanding
on the effects of as-recorded mainshockaftershock sequences in
the response of MDOF systems. For that purpose, an earthquakes
ground motion ensemble of 64 seismic sequences recorded in
accelerographic stations during the 1994 Northridge and the 1980
Mammoth Lakes earthquakes was considered in this investigation.
Specific goals of the research reported in this paper were:
(a) to investigate the correlation between frequency content
of the mainshock and main aftershock, (b) to study the effect
of aftershocks on peak and residual (permanent) displacement
demands in MDOF systems, and (c) to investigate differences in the
response due to artificial and as-recorded sequences. It should be
noted that most of the prior studies focused their attention on the
effect of seismic sequences on the level of lateral strength demands
(i.e. response modification factor) and peak displacement demands
(i.e. displacement ductility demand, inelastic displacement ratio),
but they did not provide statistical information about the effect of
aftershocks on residual (permanent) displacements at the end of
the mainshock. Residual drift demands have been identified as one
of the most important response parameters in the evaluation of the
residual capacity of damaged structures to sustain aftershocks [8]
and their evaluation play a very important role for assessment
the seismic performance of existing structures [e.g. [18,19]].
Furthermore, recent studies have noted that neglecting losses from
residual drifts can lead to significant underestimation of economic
losses [20].
Although the reported investigation provides further informa-
tion about the effects of seismic sequences on existing structures,
it should be noted that proposing an evaluation procedure to take
into account the effect of aftershocks is beyond the scope of the
research reported in this paper.
2. Existing steel frames and earthquake ground motions
considered in this study
2.1. Building frame models
Three regular three-bay frame models having three different
numbers of stories (N = 4, 8, and 12), which are representative
of exterior steel moment-resisting frames found in typical
low-to-medium height-rise existing steel office buildings, were
considered in this investigation. Fig. 1 illustrates the plan view,
the geometry and steel sections of the analyzed frames. It should
be mentioned that all frames have uniform mass distribution
and a non-uniform lateral stiffness distribution over the height.
The frame models were originally designed by Santa-Ana and
Miranda [21] using the lateral load distribution specified in the
1994 Uniform Building Code for a structure located in Zone 4 on
soil type S1 (i.e. a soil profile with either a rock-like material or
stiff or dense soil).
The buildings were analyzed using the nonlinear dynamic
analysis computer program RUAUMOKO [22]. Only half of the
building was modeled due to symmetry in the buildings plan.
The exterior frame was modeled as a two-dimensional centerline
model, assuming fixed columns, with an additional fictitious
column. The fictitious column carries the vertical (gravity) loading
 J. Ruiz-Garca, J.C. Negrete-Manriquez / Engineering Structures 33 (2011) 621634 623

TYPICAL PLANT

7.31m

FRAME ANALYSED
21.94m 7.31m

7.31m

7.315m 7.315m 7.315m

21.945m

Fig. 1. Plan view and elevation of existing steel buildings considered in this study.

from the rest of building (i.e. vertical loading carried by the
interior gravity columns) and is attached to the exterior frame
model through rigid frame elements to experience the same lateral
deformation at each floor. However, the fictitious column does not
provide the additional lateral stiffness from the interior gravity
columns. Beams and columns were modeled as frame elements
which concentrate their inelastic response in plastic hinges
located at their ends. A non-degrading bilinear momentcurvature
relationship with strain-hardening ratio equal to 2% that considers
axial load-flexural bending interaction was considered to model
the hysteretic behavior of the steel columns. The beams behavior
was modeled through a bilinear momentcurvature relationship
with strain-hardening ratio equal to 2% that includes strength
degradation due to fracture according to what has been discussed
in [23]. Flexural moment capacity for beams and columns was
determined using actual yield strength capacity of 337.8 MPa and
399.9 MPa, respectively. However, additional strength and stiffness
due to floor slab contribution in beams was neglected. Rayleigh
damping equal to 5% of critical was assigned to the first and second
modes for the 4-story frame, while this damping was applied to the
first and fourth modes for the 8- and 12-story frames. During the
analysis, local P-delta effects were included (i.e. large displacement
analysis). Main dynamic and mechanical properties of each frame
obtained from conventional modal analysis and nonlinear static
(pushover) analysis are summarized in Table 1.
2.2. Set of mainshockaftershock seismic sequences
2.2.1. Ensemble of mainshockaftershock seismic sequences
This study focused its attention on the effects of after-
shocks on the inter-story drift demands induced to existing
Table 1
Fundamental period of vibration, T1 , second-mode period of vibration, T2 , roof
yield displacement, y,roof , yield strength coefficient, Cy , and normalized modal
participation factor, 1 1,roof obtained for each frame model.
N T1 (s) T2 (s) y,roof (mm) Cy 1 1
4 1.23 0.39 160 0.32 1.22
8 1.95 0.69 360 0.25 1.31
12 2.61 0.93 470 0.18 1.31
steel frames located in regions of high-seismicity (e.g. Califor-
nia State in the United States). For this purpose, as-recorded
mainshockaftershock acceleration time histories are needed for
performing nonlinear dynamic analyses. In this study, the fol-
lowing criteria were employed for identifying and selecting
mainshockaftershock seismic sequences recorded in California,
United States, from the Pacific Earthquake Engineering (PEER)
database [17]: (a) magnitude of main aftershock event equal to
or greater than 4.0; (b) available information about the soil con-
dition, which correspond to Soil Type A, B, C or D (i.e. bedrock
and stiff soils); (c) acceleration time histories recoded on stations
placed on free field or low-height buildings where soilstructure
interaction effects are negligible; and (d) seismic sequences hav-
ing peak ground acceleration (PGA) of the mainshock horizontal
component greater than 100 cm/s2 and PGA of the aftershocks
greater than 50 cm/s2 . Under these criteria 58 seismic sequences
from two orthogonal horizontal components were identified from
the 1994 Northridge earthquake (Mw = 6.7), 30 from the 1979
Imperial Valley earthquake (M = 6.5), 6 from the 1980 Mammoth
Lakes earthquake (M = 6.2), 4 from the 1980 Livermore earth-
quake (M = 5.8), and one from the 1961 Hollister (M = 6.7) and
624 J. Ruiz-Garca, J.C. Negrete-Manriquez / Engineering Structures 33 (2011) 621634

a b

c d

Fig. 2. Examples of near-fault seismic sequences considered in this study: (a) Jensen Filter Plant station (comp. 22), (b) Rinaldi Receiving station (comp. 228), (c) Newhall
station (comp. 180), (d) Sylmar Converter Station (comp. 288).

a b

c d

Fig. 3. Velocity time-history of aftershocks recorded in accelerographic stations located near the causative fault during the 1994 Northridge earthquake: (a) Jensen Filter
Plant station (comp. 22), (b) Rinaldi Receiving station (comp. 228), (c) Newhall station (comp. 180), (d) Sylmar Converter Station (comp. 288).

the 1983 Coalinga earthquake (M = 6.4). In this investigation, re-
sults are reported from relevant observations of seismic sequences
recorded during the 1994 Northridge and the 1980 Mammoth
Lakes earthquakes. A distinct feature from previous studies is that
the catalogue contains 14 seismic sequences captured in accelero-
graphic stations located near the causative fault during the 1994
Northridge earthquake. For example, seismic sequences at Rinaldi
Receiving Station and Sylmar Converter Station located in the Los
Angeles area were considered in this study. Examples of far-field
and near-fault seismic sequences are illustrated in Fig. 2.
It should be noted that the aftershocks recorded in the near fault
region also exhibit pulse-like features in the velocity time-history
as illustrated in Fig. 3. Information about the selected as-recorded
seismic sequences can be found in Tables 2 and 3.
2.2.2. Relationship between mainshockaftershock frequency content
In order to study whether the main aftershocks accelera-
tion time-histories have similar frequency content as their corre-
sponding mainshock time-histories, the predominant period of the
ground motion (Tg ) and the bandwith ( ) were selected as a mea-
sure of the frequency content. Both frequency measures were de-
rived from the elastic velocity spectra. The predominant period of
the ground motion was defined as the period at which the maxi-
mum ordinate of a five percent damped relative velocity spectrum
occurs [24]. For illustration purposes, estimation of Tg from the ve-
locity spectra of three mainshock and main-aftershock is shown
in Fig. 4. It can clearly be seen that the mianshock and the main-
aftershock have very different predominant periods of ground mo-
tions and, as a consequence, very different frequency content.
In addition, an analogue measure of the ground motion spread
about the central period, or bandwidth, as a function of the
spectral parameters computed from the squared velocity spectra
was employed in this study [24,25], which allows defining whether
a ground motion has narrowband or broadband frequency content
around its central frequency or central period. For example, fol-
lowing the aforementioned definitions, the relationship between
Tg and for all 44 ordinary and 14 near-fault mainshock and main-
aftershock ground motions recorded during the 1994 Northridge
earthquake is shown in Fig. 5(a) and (b), while a similar plot for
the 1980 Mammoth Lakes sequences is shown in Fig. 5(c). From the
figures, it can be seen that the relationship between the frequency
content of the mainshockaftershocks is different. For the main-
shockaftershock scenarios of the 1994 Northridge earthquake:
(1) the predominant period of the mainshocks follows a linear
trend with respect to its bandwidth (e.g. Tg decreases as in-
creases), and (2) predominant periods as well as bandwidth of
main-aftershock ground motions tend to be shorter and larger,
respectively, than the predominant period of mainshock ground
motions. According with the definition postulated in [25], larger
values of the bandwidth are associated with broad band signals,
which means that aftershock ground motions are richer in high-
frequency content than mainshock ground motions.
In addition, the relationship between the predominant period
of the ground motion corresponding to the mainshock and
main-aftershock is also shown in Fig. 6(a) and (b) for the
seismic sequences recorded during the Northridge earthquake.
The sample correlation coefficient computed for both cases was
0.18 and 0.15, respectively, which means that the frequency
content of the mainshock and main-aftershock ground motions is
weakly correlated from a statistical point-of-view. From the latter
observation, it can be concluded that the simulation approach of
repeating the mainshock as an aftershock is not appropriate.
 J. Ruiz-Garca, J.C. Negrete-Manriquez / Engineering Structures 33 (2011) 621634 625

Table 2
List of mainshockaftershock seismic sequences (far-field ground motions) considered in this investigation.

Sequence Station name MODY HRMN Component Component Earthquake PGA Tg

name magnitude (cm/s2 )

GRIF19940117 LA Griffith park observatory 0117 1231 NORTHR/0141-270.at2 270 6.7 283.6 0.51 0.61
NORTHR/0141-360.at2 360 160.4 0.88 0.57
0320 2120 NORTH392/0GPN00E.at2 0 5.3 31.2 0.49 0.60
NORTH392/0GPN90W.at2 90 54.9 0.50 0.66

HOSF19940117 LA Hollywood Stor FF 0117 1231 NORTHR/PEL090.at2 90 6.7 226.9 0.73 0.51
NORTHR/PEL360.at2 360 351.4 0.85 0.58
0117 1232 NORTH001/HOLLY090.at2 90 6.1 154.8 0.15 0.65
NORTH001/HOLLY360.at2 360 162.2 0.43 0.55
0320 2120 NORTH392/HOLLY090.at2 90 5.3 107.0 0.21 0.68
NORTH392/HOLLY360.at2 360 282.9 0.25 0.61

NFAR19940117 LA N Faring Rd 0117 1231 NORTHR/FAR000.at2 0 6.7 267.7 0.62 0.59

NORTHR/FAR090.at2 90 237.3 0.61 0.54
0320 2120 NORTH392/FAR000.at2 0 5.3 103.2 0.56 0.67
NORTH392/FAR090.at2 90 119.6 0.26 0.62

OBRE19940117 LA Obregon Park 0117 1231 NORTHR/OBR090.at2 90 6.7 348.1 0.45 0.61
NORTHR/OBR360.at2 360 552.1 0.27 0.61
0117 1232 NORTH001/OBREP090.at2 90 6.1 65.0 0.19 0.61
NORTH001/OBREP360.at2 360 55.4 0.15 0.62

WOND19940117 LA Wonderland Ave 0117 1231 NORTHR/WON095.at2 95 6.7 109.8 1.14 0.50
NORTHR/WON185.at2 185 168.7 0.85 0.52
0320 2120 NORTH392/WON095.at2 95 5.3 47.3 0.26 0.61
NORTH392/WON185.at2 185 50.6 0.27 0.64

CRES19940117 La Crescenta New York 0117 1231 NORTHR/NYA090.at2 90 6.7 174.5 0.48 0.62
NORTHR/NYA180.at2 180 155.9 0.56 0.55
0320 2120 NORTH392/NYA090.at2 90 5.3 93.0 0.48 0.63
NORTH392/NYA180.at2 180 68.6 0.51 0.62

MOOR19940117 Moorpark Fire Sta 0117 1231 NORTHR/MRP090.at2 90 6.7 189.3 1.28 0.51
NORTHR/MRP180.at2 180 286.3 0.42 0.48
0117 2333 NORTH142/MPARK090.at2 90 5.9 136.8 0.91 0.53
NORTH142/MPARK180.at2 180 180.7 0.49 0.51

NORT19940117 Northridge 17645 Saticoy St 0117 1231 NORTHR/STC090.at2 90 6.7 360.9 1.31 0.47
NORTHR/STC180.at2 180 467.7 2.29 0.45
0320 2120 NORTH392/STC090.at2 90 5.3 100.2 0.78 0.61
NORTH392/STC180.at2 180 170.9 0.23 0.59

SANT19940117 Santa Monica City Hall 0117 1231 NORTHR/STM090.at2 90 6.7 865.9 2.29 0.45
NORTHR/STM360.at2 360 362.8 2.53 0.44
0320 2120 NORTH392/STM090.at2 90 5.3 100.2 0.23 0.59
NORTH392/STM360.at2 360 75.8 0.59 0.56

SIMI19940117 Simi Valley Katherine Rd 0117 1231 NORTHR/KAT000.at2 0 6.7 860.0 0.67 0.61
NORTHR/KAT090.at2 90 627.6 0.53 0.61
0320 2120 NORTH392/KAT000.at2 0 5.3 65.5 0.34 0.67
NORTH392/KAT090.at2 90 103.4 0.32 0.65

SUNV19940117 Sun Valley Roscoe Blvd 0117 1231 NORTHR/RO3000.at2 0 6.7 297.1 1.08 0.49
NORTHR/RO3090.at2 90 434.4 1.16 0.49
0320 2120 NORTH392/RO3000.at2 0 5.3 115.0 0.27 0.63
NORTH392/RO3090.at2 90 116.9 0.54 0.61

SUNL19940117 Sunland Mt Gleason Ave 0117 1231 NORTHR/GLE170.at2 170 6.7 124.5 1.13 0.49
NORTHR/GLE260.at2 260 154.0 1.00 0.56
0320 2120 NORTH392/GLE170.at2 170 5.3 86.0 0.58 0.64
NORTH392/GLE260.at2 260 154.2 0.53 0.68

(continued on next page)

626 J. Ruiz-Garca, J.C. Negrete-Manriquez / Engineering Structures 33 (2011) 621634

Table 2 (continued)

Sequence name Station name MODY HRMN Component Earthquake magnitude PGA Tg
(cm/s2 )

ARLE19940117 Arleta Nordhoff Fire Sta 0117 1231 NORTHR/ARL090.at2 6.7 337.4 1.05 0.50
NORTHR/ARL360.at2 302.1 2.43 0.44
0117 1232 NORTH001/ARLET090.at2 6.1 98.4 0.28 0.57
NORTH001/ARLET360.at2 98.7 0.27 0.61
0320 2120 NORTH392/ARLET090.at2 5.3 117.0 0.34 0.66
NORTH392/ARLET360.at2 123.1 0.23 0.65

BEVE19940117 Beverly Hills 12520 Mulhol 0117 1231 NORTHR/MU2035.at2 6.7 604.9 0.26 0.51
NORTHR/MU2125.at2 435.8 0.26 0.58
0320 2120 NORTH392/MU2035.at2 5.3 175.6 0.28 0.63
NORTH392/MU2125.at2 161.0 0.25 0.65

BIGT19940117 Big Tujunga, Angeles Nat F 0117 1231 NORTHR/TUJ262.at2 6.7 160.1 0.29 0.62
NORTHR/TUJ352.at2 240.4 0.65 0.59
0320 2120 NORTH392/TUJ262.at2 5.3 131.6 0.26 0.63
NORTH392/TUJ352.at2 103.8 0.28 0.62

BURB19940117 Burbank Howard Rd. 0117 1231 NORTHR/HOW060.at2 6.7 117.4 0.64 0.56
NORTHR/HOW330.at2 160.1 0.28 0.61
0320 2120 NORTH392/HOW060.at2 5.3 75.4 0.27 0.61
NORTH392/HOW330.at2 67.2 0.44 0.63

CAST19940117 Castaic Old Ridge Route 0117 1231 NORTHR/ORR090.at2 6.7 557.3 0.82 0.47
NORTHR/ORR360.at2 504.3 0.95 0.51
0117 2333 NORTH142/CASTA090.at2 5.9 135.7 0.39 0.62
NORTH142/CASTA360.at2 116.6 0.31 0.59
0117 0043 NORTH151/CASTA090.at2 5.1 103.1 0.39 0.62
NORTH151/CASTA360.at2 59.1 0.31 0.59

DOWN19940117 Downey Co Maint Bldg 0117 1231 NORTHR/DWN090.at2 6.7 155.2 0.62 0.52
NORTHR/DWN360.at2 225.5 0.37 0.56
0117 1232 NORTH001/DOWNC090.at2 6.1 50.4 0.28 0.57
NORTH001/DOWNC180.at2 33.7 0.25 0.59

HOLL19940117 Hollywood Willoughby Ave 0117 1231 NORTHR/WIL090.at2 6.7 133.7 0.75 0.48
NORTHR/WIL180.at2 240.8 1.08 0.52
0320 2120 NORTH392/WIL090.at2 5.3 68.0 0.24 0.65
NORTH392/WIL180.at2 99.1 0.56 0.61

BALD19940117 LA Baldwin Hills 0117 1231 NORTHR/BLD090.at2 6.7 234.2 1.98 0.43
NORTHR/BLD360.at2 164.5 1.63 0.47
0117 1232 NORTH001/BALDW090.at2 6.1 111.4 0.29 0.60
NORTH001/BALDW360.at2 48.2 0.52 0.61
0320 2120 NORTH392/BALDW090.at2 5.3 61.1 0.53 0.61
NORTH392/BALDW180.at2 53.0 0.46 0.52

CENT19940117 LA Century City CC North 0117 1231 NORTHR/CCN090.at2 6.7 250.7 1.60 0.43
NORTHR/CCN360.at2 217.4 1.06 0.49
0117 1232 NORTH001/CNTYC090.at2 6.1 109.1 0.30 0.65
NORTH001/CNTYC360.at2 58.4 0.61 0.59
0320 2120 NORTH392/CNTYC090.at2 5.3 159.0 0.50 0.66
NORTH392/CNTYC360.at2 84.3 0.58 0.62

CITY19940117 LA City Terrace 0117 1231 NORTHR/LAC090.at2 6.7 258.0 0.32 0.57
NORTHR/LAC180.at2 310.2 0.41 0.63
0117 1232 NORTH001/CTYTE090.at2 6.1 51.3 0.24 0.65
NORTH001/CTYTE180.at2 112.3 0.19 0.57
0320 2120 NORTH392/CTYTE090.at2 5.3 32.6 0.24 0.62
NORTH392/CTYTE180.at2 60.2 0.24 0.61
 J. Ruiz-Garca, J.C. Negrete-Manriquez / Engineering Structures 33 (2011) 621634 627

Table 3
List of mainshockaftershock seismic sequences (near-fault) considered in this investigation.
Sequence name Station name MODY HRMN Peer file name Earthquake magnitude PGA Tg
(cm/s2 )

JEFP19940117 Jensen Filter Plant 0117 1231 NORTHR/JEN022.at2 6.7 559.6 2.94 0.33
NORTHR/JEN292.at2 1004.0 1.09 0.49
0117 0043 NORTH151/00BN22E.at2 5.1 47.4 0.49 0.56
NORTH151/00BW22N.at2 47.4 0.49 0.56
0320 2120 NORTH392/00BN22E.at2 5.3 208.1 0.60 0.57
NORTH392/00BW22N.at2 250.4 0.42 0.62

JEFG19940117 Jensen Filter Plant Generator 0117 1231 NORTHR/0655-022.at2 6.7 559.6 2.94 0.33
NORTHR/0655-292.at2 1003.7 1.04 0.49
0117 0043 NORTH151/00GN22E.at2 5.1 70.3 0.30 0.64
NORTH151/00GW22N.at2 56.7 0.66 0.61
0320 2120 NORTH392/00GN22E.at2 5.3 154.1 0.43 0.57
NORTH392/00GW22N.at2 297.3 0.42 0.61

NEWH19940117 Newhall Fire Station 0117 1231 NORTHR/NWH090.at2 6.7 571.7 1.29 0.52
NORTHR/NWH360.at2 578.3 1.27 0.50
0117 1232 NORTH001/NEWHA090.at2 6.1 35.8 0.50 0.62
NORTH001/NEWHA360.at2 42.7 0.40 0.65
0117 1241 NORTH009/NEWHA090.at2 5.2 104.9 0.46 0.58
NORTH009/NEWHA180.at2 201.5 0.63 0.56
0320 2120 NORTH392/NEWHA090.at2 5.3 53.9 0.24 0.65
NORTH392/NEWHA360.at2 59.1 0.65 0.61

PACO19940117 Pacoima Kagel Canyon 0117 1231 NORTHR/PKC090.at2 6.7 295.2 0.86 0.47
NORTHR/PKC360.at2 424.3 0.65 0.53
0117 1232 NORTH001/PACOI090.at2 6.1 63.8 0.29 0.65
NORTH001/PACOI360.at2 78.4 0.24 0.61
0117 1241 NORTH009/PACOI090.at2 5.2 34.7 0.36 0.52
NORTH009/PACOI360.at2 52.4 0.93 0.57
0117 0043 NORTH151/PACOI090.at2 5.1 53.0 0.82 0.57
NORTH151/PACOI360.at2 21.8 0.52 0.53
0320 2120 NORTH392/PACOI090.at2 5.3 149.1 0.31 0.61
NORTH392/PACOI360.at2 223.5 0.68 0.64

RINA19940117 Rinaldi Receiving Sta 0117 1231 NORTHR/RRS228.at2 6.7 809.2 1.05 0.47
NORTHR/RRS318.at2 477.1 2.34 0.40
0320 2120 NORTH392/E-RRS228.at2 5.3 639.0 0.41 0.63
NORTH392/E-RRS318.at2 421.9 0.40 0.61

SYLM19940117 Sylmar Converter Sta East 0117 1231 NORTHR/SCE018.at2 6.7 812.2 0.83 0.41
NORTHR/SCE288.at2 483.5 2.20 0.39
0320 2120 NORTH392/E-SCE018.at2 5.3 154.8 0.88 0.63
NORTH392/E-SCE288.at2 122.6 0.22 0.65

TARZ19940117 Tarzana Cedar Hill A 0117 1231 NORTHR/TAR090.at2 6.7 1744.8 0.60 0.56
NORTHR/TAR360.at2 971.1 0.74 0.56
0117 1241 NORTH009/TARZA090.at2 5.2 67.6 0.35 0.64
NORTH009/TARZA360.at2 54.8 0.37 0.65
0320 2120 NORTH392/TARZA090.at2 5.3 365.4 0.22 0.67
NORTH392/TARZA360.at2 302.6 0.25 0.65

3. Response under seismic sequences
3.1. Response under individual seismic sequences
At a first stage, the response of the 4-story frame model under
selected near-fault mainshockaftershock seismic sequences is
examined and, therefore, displacement time-history results are
shown in Fig. 7. It can be seen that the response of this frame model
is dominated by the mainshock, which means that the aftershocks
do not increase peak displacement demands, and furthermore,
aftershocks do not increase permanent displacements triggered
at the end of the mainshock. Similar observations were made for
the frame model under far-field ground motions from the 1994
Northridge earthquake.
The aforementioned observation seems opposite to the results
discussed in previous studies. However, it should be recall that
most of the previous studies have developed investigations on
the effect of aftershocks in the response of structures employing
artificial seismic sequences [e.g. [713,15]]. Therefore, it is of
interest to examine the response of frame models under as-
recorded and artificial seismic sequences. For this purpose, let
us consider the as-recorded mainshockaftershock acceleration
time-history recorded in the Rinaldi Receiving Station (component
228) during the 1994 Northridge earthquake (Fig. 2(b)). This
628 J. Ruiz-Garca, J.C. Negrete-Manriquez / Engineering Structures 33 (2011) 621634

a b c

Fig. 4. Estimation of predominant period of the ground motion from the velocity spectra of typical mainshock and main-aftershock near-fault ground motions considered
in this study: (a) Tarzana (comp. 090); (b) Rinaldi (comp. 228); (c) Jensen Filter Plant Generator (comp. 022).

a b

Fig. 5. Relationship between Tg and for mainshocks and main aftershocks: (a) far-field ground motions (1994 Northridge earthquake), (b) near-fault ground motions
(1994 Northridge earthquake), (c) far-field ground motions (1980 Mammoth Lakes earthquake).

a b

Fig. 6. Relationship of Tg for mainshock and Tg for aftershocks: (a) far-field ground motions (1994 Northridge earthquake), (b) near-fault ground motions (1994 Northridge
earthquake).

Fig. 7. Response (first-story displacement time-history) of the 4-story frame model under selected near-fault seismic sequences.
 J. Ruiz-Garca, J.C. Negrete-Manriquez / Engineering Structures 33 (2011) 621634
a b
Fig. 8. Seismic sequences in the near-fault environment (Rinaldi Receiving station, comp. 228): (a) as-recorded; (b) back-to-back case; (c) randomized case.
a at the end of the excitation in both frames, mainly in the 4-story
b the period of vibration of the damaged frames. On the other
Fig. 9. Displacement time-history response of frame models under as-recorded and
artificial seismic sequences from the Rinaldi Receiving station (comp. 228): (a) 4-
story frame; and (b) 12-story frame.
seismic sequence is chosen since the main aftershock has the
largest PGA among the recorded aftershocks in the catalogue and
it has a (PGA)M /(PGA)A ratio equal to 0.79 (see Fig. 2(b)). In
addition, artificial seismic sequences are generated by repeating
the mainshock as the aftershock (i.e. back-to-back case) and using
the mainshock acceleration time-history recorded in the Sylmar
Converter station as an aftershock (i.e. randomized case). Fig. 8
shows the as-recorded and artificial seismic sequences. It can
clearly be seen that the artificial seismic sequences have very
different ground motion features than the as-recorded seismic
sequence.
Therefore, each frame was subjected to the as-recorded seismic
sequence and the two artificial sequences. For instance, the
displacement time-history obtained from the first story of the
4- and 12-story frame models is shown in Fig. 9. From the figure, it
can be seen that the mainshock triggers permanent displacement
629
frame. This observation could be explained since the calculated
predominant period of the mainshock ground motion is 1.05 s,
which is close than the first-mode period of vibration of the
4-story frame (T1 = 1.23 s, Tg /T1 = 0.33). However, the as-
recorded aftershock does not increase either peak and permanent
displacement demands in both frames, which might be explained
since the predominant period of the aftershock (Tg = 0.41 s) is
shorter than either the fundamental period of vibration of both
frames (i.e. undamaged state at the end of the mainshock) or
hand, it should be noted that artificial seismic sequences lead to
larger permanent displacements of each frame at the end of the
artificial aftershock than the as-recorded sequence. In both cases,
the artificial sequence defined by the back-to-back approach leads
to the largest permanent displacement.
It is of particular interest to further investigate the effect
of the frequency content of the aftershock in the building
response. However, all the near-fault aftershocks in the ground
motion database have Tg s shorter than the buildings fundamental
period of vibration. Then, artificial aftershocks having different
frequency content (i.e. pulse period) were generated through
a velocity pulse model proposed by Menun and Fu [26]. For
example, Fig. 10(a) illustrates the recorded aftershock ground
velocity time-history recorded at the Rinaldi Receiving station
and the fitted velocity pulse model, while Fig. 10(b) shows a
similar comparison for the aftershock acceleration time-history.
As discussed by Kalkan and Kunnath [27], acceleration time-
histories developed from velocity pulse models do not necessarily
match recorded acceleration time-histories. Using the as-recorded
mainshock and the acceleration time-history with acceleration
time-history obtained from the velocity pulse model, it was
first verified that this artificial sequence could reproduce the
displacement time-history obtained from the real sequence.
Next, the velocity pulse model was calibrated to have different
pulse periods (i.e. frequency content) while keeping the same
peak ground acceleration of the as-recorded aftershock. The
displacement time-history recorded in the first story of the 4-
story frame model when subjected to as-recorded mainshock-
pulse type aftershocks sequences is shown in Fig. 11. From the
630 J. Ruiz-Garca, J.C. Negrete-Manriquez / Engineering Structures 33 (2011) 621634

a drift demand (i.e. relative inter-story displacement normalized

b specifies permanent inter-story drift thresholds of 0.7% and 5% for
Fig. 10. Comparison of as-recorded ground velocity and acceleration of main-
aftershock recorded at the Rinaldi Receiving station with a pulse model proposed
in [26].
figure, it can clearly be observed the effect of the aftershocks
frequency content in permanent displacements at the end of the
sequence.
3.2. Response of frame models under as-recorded far-field and near-
fault seismic sequences
In order to study the influence of mainshockaftershock
sequences in the response of existing steel frames, a series of
nonlinear dynamic analyses were carried out for each frame
model when subjected to the as-recorded seismic sequences.
Later, individual results were statistically processed to obtain the
central tendency and dispersion of peak (transient) inter-story
with respect to the story height), IDR, and residual inter-story
drift demand (i.e. relative inter-story permanent displacement at
the end of the earthquakes excitation normalized with respect
to the story height), RIDR. In order to provide a context of the
frames performance, FEMA 356 [1] recommendations specify
inter-story drift index thresholds of 0.7%, 2.5% and 5% for the
immediate occupancy (IO), life safety (LS) and collapse prevention
(CP) performance levels, respectively. In addition, FEMA 356 [1]
life safety and collapse prevention, respectively.
The distribution over the height of IDR computed from the
set of seismic sequences recorded during the 1980 Mammoth
Lakes earthquakes for the three frames models is shown in
Fig. 12. It can be seen that the mainshocks trigger inter-story drift
demands shorter than the inter-story drift thresholds linked to
the IO performance level. However, IDRs are increased due to the
aftershocks, but only the 4-story frame exceeds the IDR associated
to IO performance level. It should be noted that the increment in
IDR due to aftershocks depends on the period of vibration, number
of stories and the story height. In addition, it was observed that
the aftershocks increased the level of RIDR at the end of the
mainshocks, but the level of RIDR was negligible below the IO
performance level.
Fig. 13 shows the height-wise distribution of IDR obtained
for each frame under all 44 far-field and all 14 near-fault as-
recorded seismic sequences captured during the 1994 Northridge
earthquake. As can be expected, near-fault earthquake ground
motions induce larger IDRs than the far-field ground motions
which tends to concentrate in the ground story since it was
anticipated that the frames could develop a ground soft story
mechanism. From the figure, it can be observed that the 4- and
8-story frames exceed the drift thresholds associated to the life-
safety level under near-fault mainshocks, but all frames experience
IDRs smaller than those prescribed for the IO performance level
under far-field mainshock ground motions. Particularly, it can be

Fig. 11. Displacement time-history response of the 4-story frame model under as-recorded mainshock and artificial pulse-type aftershocks.

a b c

Fig. 12. Height-wise distribution of peak (transient) inter-story drift demand computed from 6 seismic sequences recorded during the 1980 Mammoth Lakes earthquakes:
(a) 4-story frame; (b) 8-story frame; and (c) 12-story frame.
 J. Ruiz-Garca, J.C. Negrete-Manriquez / Engineering Structures 33 (2011) 621634
a b
Fig. 13. Height-wise distribution of median peak (transient) inter-story drift demand computed from far-field and near-fault seismic sequences recorded during the 1994
Northridge earthquakes: (a) 4-story frame; (b) 8-story frame; and (c) 12-story frame.
z/H z/H
a b
RIDR [%]
Fig. 14. Height-wise distribution of median residual (permanent) inter-story drift demand under all 14 near-fault seismic sequences: (a) 4-story frame; (b) 8-story frame;
and (c) 12-story frame.
seen that response of the 12-story frame is more influenced by
higher modes under near-fault ground motions than under far-
field motions. Regarding the effect of aftershocks, it can also be
seen that the mainshock controls the response regardless of the
type of ground motion (i.e. the aftershocks do not increase the level
of drift demands).
An important task in this investigation was to investigate
whether the aftershocks could increase residual inter-story drift
demands at the end of the mainshock. Therefore, Fig. 14 shows the
distribution over the height of RIDR corresponding to frame models
under the near-fault ground motion sequences. It can be observed
that the 8-story frame experiences the largest RIDR, which is
in excess of the permanent inter-story drift threshold for life-
safety performance level. However, permanent inter-story drift
demands are not increased as a consequence of the as-recorded
aftershocks. This observation could be explained since all the as-
recorded aftershocks have predominant periods shorter than about
1.0 s, which means that the predominant periods of the aftershock
ground motions are smaller than the first-mode period of vibration
and, as expected, much smaller than the period of vibration
of the damaged building (i.e. if the frame experienced non-
recoverable deformations). The aforementioned observation is
very important since permanent drift demands could be related to
the residual capacity of the building [8]. A similar statistical study
was conducted using the as-recorded far-field seismic sequences.
Unlike previous results for near-fault sequences, permanent drift
demands triggered by far-field sequences were very small, and
they did not increase as a consequence of the aftershocks.
631
c
z/H
c
RIDR [%] RIDR [%]
3.3. Response of frame models under artificial seismic sequences
It should be recognized that previous results, mainly for peak
inter-story drift demand, are opposite to some of the prior stud-
ies. However, recall that with exception of the studies by Ruiz-
Garca et al. [14] and Hatzeorgiou and Liolios [15], preceding stud-
ies employed artificial seismic sequences, either from a repeated or
randomized approach. Furthermore, earthquake field reconnais-
sance after the 1994 Northridge earthquake did not report dam-
age in steel buildings as a consequence of aftershocks [3,4]. As
shown in the previous section, the repeated and randomized ap-
proach could lead to the conclusion that aftershocks consistently
increase peak and permanent displacement demands from the
mainshock. Therefore, it is of interest to investigate the level of
overestimation in drift demands that artificial seismic sequences
generated from the repeated and randomized approach induce to
the frames. Fig. 15 shows a comparison of the height-wise distribu-
tion of IDR for all frames computed from the as-recorded and artifi-
cial near-fault seismic sequences. Results are presented in terms of
mean-plus-one-standard deviation in order to include the record-
to-record variability. As can be expected, it can be seen that artifi-
cial sequences trigger larger maximum IDR over all stories, while
the level of overestimation of IDR depends on the specific story
and the period of vibration of the frame. For instance, IDRs in the
first story of each frame computed from the repeated approach are
larger about 13%, 24% and 35% than the corresponding IDRs calcu-
lated from as-recorded sequences.
An important finding for probabilistic assessment of buildings
under a mainshockaftershock scenario is that the record-to-
record variability, measured by the logarithmic standard deviation
632 J. Ruiz-Garca, J.C. Negrete-Manriquez / Engineering Structures 33 (2011) 621634

z/H z/H z/H

a b c

Fig. 15. Height-wise distribution of peak (transient) inter-story drift demand under as-recorded and artificial (repeated approach) near-fault seismic sequences: (a) 4-story
frame; (b) 8-story frame; and (c) 12-story frame.

z/H
a z/H
b z/H
c

Fig. 16. Height-wise distribution of log standard deviation of peak (transient) inter-story drift demand under as-recorded and artificial (repeated approach) near-fault
seismic sequences: (a) 4-story frame; (b) 8-story frame; and (c) 12-story frame.

a z/H b z/H c z/H

Fig. 17. Height-wise distribution of mean plus one standard deviation of peak (transient) inter-story drift demand under as-recorded and artificial (repeated approach)
far-field seismic sequences: (a) 4-story frame; (b) 8-story frame; and (c) 12-story frame.

of IDR, computed from the artificial sequences is, in general, larger
than that computed from the as-recorded sequences. For example,
height-wise distribution of logarithmic standard deviation of IDR
computed from the near-fault sequences is shown in Fig. 16. This
observation has impact while computing fragility curves of peak
interstory drift demands and, as a consequence, in seismic hazard
curves of peak-interstory drift demand.
A comparison of IDRs triggered by as-recorded and artificial
(repeated) far-field seismic sequences is illustrated in Fig. 17. Un-
der this seismic environment, the artificial mainshockaftershock
sequences do not trigger significantly different mean-plus-one-
standard deviation IDRs than the as-recorded seismic sequences.
For example, IDRs in the first story computed from the repeated
approach are larger about 13%, 33% and 6% than the corresponding
IDRs calculated from as-recorded sequences.
Finally, a comparison of mean-plus-one-standard deviation
RIDRs computed from as-recorded and artificial (repeated)
near-fault sequences is shown in Fig. 18. It can clearly be seen that
 J. Ruiz-Garca, J.C. Negrete-Manriquez / Engineering Structures 33 (2011) 621634
a z/H
b z/H
RIDR [%]
Fig. 18. Height-wise distribution of mean plus one standard deviation of residual (permanent) inter-story drift demand under as-recorded and artificial (repeated approach)
near-fault seismic sequences: (a) 4-story frame; (b) 8-story frame; and (c) 12-story frame.
the use of artificial seismic sequences tends to overestimate the
amplitude of residual drift demands. The level of overestimation
depends on the number of stories and the story height.
4. Summary and conclusions
This paper has summarized the results of an analytical study
aimed at providing further understanding on the influence of
aftershocks on drift demands in regular existing moment-resisting
frame buildings. This investigation focused on investigating
whether as-recorded aftershocks could increase peak (transient)
and (residual) permanent drift demands in frame models with
different number of stories. For that purpose, a set of 58
as-recorded mainshockaftershock earthquake ground motions
recorded in far-field and near-fault accelerographic stations during
the January 17, 1994, Northridge earthquake and a set 6 seismic
sequences recorded during the 1980 Mammoth Lakes earthquakes
were considered in this study. The following conclusions are drawn
from this investigation:
The frequency content, measured by the predominant period of
the ground motion, Tg , and bandwidth, of the mainshock and
the main aftershock of the far-field and near-fault seismic se-
quences considered in this investigation are weakly correlated
(from a statistical point of view). In particular, the main after-
shocks recorded during the 1994 Northridge earthquake have
shorter Tg than their corresponding mainshock. Thus, there is
no evidence that supports simulating seismic sequences using
the mainshock as seed for reproducing the aftershock (i.e. seis-
mic sequences simulated as back-to-back mainshocks).
Unlike previous results based on artificial seismic sequences,
it was found that as-recorded aftershocks do not significantly
increase peak and permanent drift demands of existing
steel frames, which is particularly true for the seismic
sequences recorded during the 1994 Northridge earthquake.
This observation can be explained since the frequency content
of the aftershock is shorter, and in some cases much smaller,
that the frequency of the structure at the end of the mainshock.
Artificial seismic sequences lead to overestimation of the
maximum lateral drift demands as well as to the record-
to-record variability. The level of overestimation depends on
the approach for developing artificial sequences (repeated or
randomized approach).
Based on the results of this investigation, it is suggested that
the effect of aftershocks during performance-based assessment
of existing structures should be taken into account by using real
(i.e. as-recorded) mainshockaftershock seismic sequences instead
633
c z/H
RIDR [%] RIDR [%]
of artificial sequences as well as site-specific seismic scenarios
due to the particular dependency on ground motion features
(e.g. frequency content) of mainshockaftershock ground motions.
Acknowledgements
The authors would like to express their gratitude to Universidad
Michoacana de San Nicols de Hidalgo and the National Council
for Science and Technology (CONACYT) in Mxico for the support
provided to develop the research reported in this paper.
References
[1] Federal emergency management agency. Prestandard and commentary for the
seismic rehabilitation of buildings. Report FEMA 356 2000. Washington (DC).
[2] Rosenblueth E, Meli R. The 1985 Mexico earthquake: causes and effects in
Mexico City. Concr Int (ACI) 1986;8(5):2334.
[3] Anderson JC, Gourley BC, Green M, Hajjar JF, Johnston R, Leon RT, Sullivan DP,
Partridge JE. Case studies of steel moment frame building performance in the
Northridge earthquake of January 17, 1994. Report no. SAC 95-07. SAC Joint
Venture 1997. Richmond (California).
[4] Bertero VV, Anderson JC, Krawinkler H. Performance of steel building
structures during the Northridge earthquake. Report EERC 94/09. University
of California at Berkeley; 1994.
[5] Mahin SA. Effects of duration and aftershocks on inelastic design earthquakes.
In: Proceedings of the seventh world conference on earthquake engineering.
vol. 5. 1980. p. 6779.
[6] Sunasaka Y, Kiremidjian A. A method for structural safety evaluation under
mainshockaftershock earthquake sequences. Report no. 105. The John A.
Blume Earthquake Engineering Center, Stanford University; 1993.
[7] Amadio C, Fragiacomo M, Rajgelj S. The effects of repeated earthquake ground
motions on the non-linear response of SDOF systems. Earthq Eng Struct Dyn
2003;32:291308.
[8] Luco N, Bazzurro P, Cornell CA. Dynamic versus static computation of the
residual capacity of mainshock-damaged building to withstand an aftershock.
In: Proceedings 13th. world conference on earthquake engineering. Paper no.
2405. 2004.
[9] Hatzigeorgiou GD, Beskos DE. Inelastic displacement ratios for SDOF structures
subjected to repeated earthquakes. Eng Struct 2009;31:274455.
[10] Hatzigeorgiou GD. Ductility demand spectra for multiple near- and far-fault
earthquakes. Soil Dyn Earthq Eng 2010;30:17083.
[11] Fragiacomo M, Amadio C, Macorini L. Seismic response of steel frames under
repeated earthquake ground motions. Eng Struct 2004;26:202135.
[12] Lee K, Foutch DA. Performance evaluation of damaged steel frame buildings
subjected to seismic loads. J Struct Eng 2004;130:58899.
[13] Li Q, Ellingwood BR. Performance evaluation and damage assessment of steel
frame buildings under main shockaftershock sequences. Earthq Eng Struct
Dyn 2007;36:40527.
[14] Ruiz-Garca J, Moreno JY, Maldonado IA. Evaluation of existing Mexican high-
way bridges under mainshockaftershock seismic sequences. In: Proceedings
of the 14th world conference on earthquake engineering. Paper 05-02-0090.
2008.
[15] Hatzigeorgiou GD, Liolios AA. Nonlinear behaviour of RC frames under
repeated strong ground motions. Soil Dyn Earthq Eng 2010;30:101025.
[16] Somerville P, Smith N, Puntamurthula S, Sun J. Development of ground
motion time histories for phase 2 of the FEMA/SAC steel project. SAC
background document no. SAC/BD-97/04. SAC joint venture 1997. Richmond
(California).
[17] Pacific Earthquake Engineering Research Center PEER Strong Motion Database.
www/http://peer.berkeley.edu/nga/. Last access: 07/20/2010.
634 J. Ruiz-Garca, J.C. Negrete-Manriquez / Engineering Structures 33 (2011) 621634
[18] Ruiz-Garca J, Miranda E. Performance-based assessment of existing structures
accounting for residual displacements. Report no. 153. Stanford (Calif.): The
John A. Blume Earthquake Engrg Ctr, Stanford Univ.; 2005.
[19] Ruiz-Garca J, Miranda E. Probabilistic estimation of residual drift demands
for seismic assessment of multi-story framed buildings. Eng Struct 2010;32:
1120.
[20] Miranda E, Ramirez M. Influence of residual displacements on building loss
estimation. In: Proceeding of the 9USNCEE. Paper no. 1223. 2010.
[21] Santa-Ana PR, Miranda E. Strength-reduction factors for multi-degree-of-
freedom systems. In: 12th. world conf. on earth. engrg. Paper 1446. 2000.
[22] Carr AJ. RUAUMOKO-inelastic dynamic analysis program. Users manual.
Christchurch (New Zealand): Dept. of Civil Engineering, University of
Canterbury; 2004.
[23] Filiatrault A, Tremblay R, Wanitkorkul A. Performance evaluation of pas-
sive damping systems for the seismic retrofit of steel moment-resisting
frames subjected to near-field ground motions. Earthq Spectra 2001;17:
427456.
[24] Miranda E. Evaluation of site-dependent inelastic seismic design spectra.
J Struct Eng ASCE 1993;119:131938.
[25] Vanmarcke EH. Properties of spectral moments with applications to random
vibrations. J Eng Mech ASCE 1972;98:42546.
[26] Menun C, Fu Q. An analytical model for near-fault ground motions and the
response of SDOF systems. In: Proceedings of the 7th US national conference
on earthquake engineering. 2002.
[27] Kalkan E, Kunnath SK. Effects of fling step and forward directivity on seismic
response of buildings. Earthq Spectra 2006;22:36790.